Liqun/fasttrackmain 9.26 (#5625)

### Description
fasttrack to current main as of 9.26 to get all PRs for 1.15.0 release

### Motivation and Context
there are many PRs in main branch that are need for 1.15.0 release so
just to a fast track merge

---------

Signed-off-by: Swopper050 <bram@witoos.nl>
Signed-off-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Signed-off-by: Justin Chu <justinchu@microsoft.com>
Signed-off-by: Justin Chu <justinchuby@users.noreply.github.com>
Signed-off-by: Chun-Wei Chen <jacky82226@gmail.com>
Signed-off-by: Liqun Fu <liqfu@microsoft.com>
Signed-off-by: Ganesan Ramalingam <grama@microsoft.com>
Signed-off-by: liqun Fu <liqun.fu@microsoft.com>
Co-authored-by: Bram <b.dewit@applyai.nl>
Co-authored-by: Swopper050 <bram@witoos.nl>
Co-authored-by: jcwchen <jcwchen@users.noreply.github.com>
Co-authored-by: Chun-Wei Chen <jacky82226@gmail.com>
Co-authored-by: Justin Chu <justinchuby@users.noreply.github.com>
Co-authored-by: G. Ramalingam <grama@microsoft.com>

.azure-pipelines/MacOS-CI.yml

@@ -82,7 +82,7 @@ jobs:
       python -m pip install onnxruntime
       export ORT_MAX_IR_SUPPORTED_VERSION=8
       export ORT_MAX_ML_OPSET_SUPPORTED_VERSION=3
-      export ORT_MAX_ONNX_OPSET_SUPPORTED_VERSION=18
+      export ORT_MAX_ONNX_OPSET_SUPPORTED_VERSION=19
       pytest
       if [ $? -ne 0 ]; then
         echo "pytest failed when testing onnx with onnxruntime"
@@ -118,5 +118,11 @@ jobs:
         echo "git diff for test generation without pillow returned failures. Please check updated node test files"
         exit 1
       fi
-
+      # Internal Protobuf won't have other untrack files like protobuf/
+      if [ '$(protobuf_type)' == 'Internal' ]; then
+        if [[ $(git ls-files --others --exclude-standard) ]]; then
+          echo "Some test-generated files not included in the PR. Did you forget to add any test-generated files?"
+          exit 1
+        fi
+      fi
     displayName: 'Run ONNX Tests'

.github/workflows/release_win.yml

@@ -72,7 +72,7 @@ jobs:
         if ('${{ github.event_name }}' -eq 'schedule') {
           echo "Build weekly PyPI package"
           (Get-Content -Path 'pyproject.toml') | ForEach-Object { $_ -replace 'name = "onnx"', 'name = "onnx-weekly"' } | Set-Content -Path 'pyproject.toml'
-          set ONNX_PREVIEW_BUILD=1
+          $Env:ONNX_PREVIEW_BUILD=1
         }
         python -m build --wheel
         Get-ChildItem -Path dist/*.whl | foreach {python -m pip install --upgrade $_.fullname}

docs/AddNewOp.md

@@ -74,7 +74,10 @@ Once the criteria of proposing new operator/function has been satisfied, you wil
     1. The testing examples will be extracted to the doc.
     2. We also generate binary data for it.
     3. Example: [onnx/backend/test/case/node/abs.py](/onnx/backend/test/case/node/abs.py)
-5. Add at least one automatic upgrade test for your operator in [onnx/test/automatic_upgrade_test.py](/onnx/test/automatic_upgrade_test.py) using `_test_op_upgrade`. These tests create a given operator at a given opset version (usually the version the operator was introduced in) and test that the version converter is able to convert them to the highest available version. So for a new operator `_test_op_upgrade` will not test anything, but as soon as the operator gets updated in a future opset the test will automatically become nontrivial.
+5. Write upgrade and downgrade tests:
+    1. Add at least one automatic upgrade test for your operator in [onnx/test/automatic_upgrade_test.py](/onnx/test/automatic_upgrade_test.py) using `_test_op_upgrade`. These tests create a given operator at a given opset version (usually the version the operator was introduced in) and test that the version converter is able to convert them to the highest available version. So for a new operator `_test_op_upgrade` will not test anything, but as soon as the operator gets updated in a future opset the test will automatically become nontrivial.
+    2. Similarly add at least one automatic downgrade test for your operator in [onnx/test/automatic_downgrade_test.py](/onnx/test/automatic_downgrade_test.py) using `_test_op_downgrade`. Specifying the current version so that once the op is updated at a higher opset version the test will ensure downward conversion is validated.
+
 6. Update the documentation and generate the test data.
     1. Running [the script](/tools/update_doc.sh). If you have files under `onnx/backend/test/data/node` which cannot be generated by the scripts from `onnx/backend/test/case/node`, please further use `python onnx/backend/test/cmd_tools.py generate-data --clean` to cleanup the directory and only preserve needed test data.
 to update the doc and generate the test data.

docs/Changelog.md

@@ -1579,7 +1579,7 @@ This version of the operator has been available since version 1 of the default O
 
 <dl>
 <dt><tt>cond</tt> : B</dt>
-<dd>Condition for the if</dd>
+<dd>Condition for the if. The tensor must contain a single element.</dd>
 </dl>
 
 #### Outputs (1 - &#8734;)
@@ -10985,10 +10985,10 @@ This version of the operator has been available since version 11 of the default
   Outputs Scale, ZeroPoint and Quantized Input for a given FP32 Input.
   Scale is calculated as:
   ```
-  y_scale = (max(x) - min(x))/(qmax - qmin)
+  y_scale = (maximum(0, max(x)) - minimum(0, min(x))) / (qmax - qmin)
   ```
 
-  * where qmax and qmin are max and min values for quantization range .i.e [0, 255] in case of uint8
+  * where qmax and qmin are max and min values for quantization range i.e. [0, 255] in case of uint8
   * data range is adjusted to include 0.
 
   Zero point is calculated as:
@@ -11525,7 +11525,7 @@ This version of the operator has been available since version 11 of the default
 
 <dl>
 <dt><tt>cond</tt> : B</dt>
-<dd>Condition for the if</dd>
+<dd>Condition for the if. The tensor must contain a single element.</dd>
 </dl>
 
 #### Outputs (1 - &#8734;)
@@ -16162,7 +16162,7 @@ This version of the operator has been available since version 13 of the default
 
 <dl>
 <dt><tt>cond</tt> : B</dt>
-<dd>Condition for the if</dd>
+<dd>Condition for the if. The tensor must contain a single element.</dd>
 </dl>
 
 #### Outputs (1 - &#8734;)
@@ -20057,7 +20057,7 @@ This version of the operator has been available since version 16 of the default
 
 <dl>
 <dt><tt>cond</tt> : B</dt>
-<dd>Condition for the if</dd>
+<dd>Condition for the if. The tensor must contain a single element.</dd>
 </dl>
 
 #### Outputs (1 - &#8734;)
@@ -23069,7 +23069,7 @@ This version of the operator has been available since version 19 of the default
 
 <dl>
 <dt><tt>cond</tt> : B</dt>
-<dd>Condition for the if</dd>
+<dd>Condition for the if. The tensor must contain a single element.</dd>
 </dl>
 
 #### Outputs (1 - &#8734;)

docs/Operators.md

@@ -7977,10 +7977,10 @@ expect(
   Outputs Scale, ZeroPoint and Quantized Input for a given FP32 Input.
   Scale is calculated as:
   ```
-  y_scale = (max(x) - min(x))/(qmax - qmin)
+  y_scale = (maximum(0, max(x)) - minimum(0, min(x))) / (qmax - qmin)
   ```
 
-  * where qmax and qmin are max and min values for quantization range .i.e [0, 255] in case of uint8
+  * where qmax and qmin are max and min values for quantization range i.e. [0, 255] in case of uint8
   * data range is adjusted to include 0.
 
   Zero point is calculated as:
@@ -11905,7 +11905,7 @@ Other versions of this operator: <a href="Changelog.md#If-1">1</a>, <a href="Cha
 
 <dl>
 <dt><tt>cond</tt> : B</dt>
-<dd>Condition for the if</dd>
+<dd>Condition for the if. The tensor must contain a single element.</dd>
 </dl>
 
 #### Outputs (1 - &#8734;)

onnx/defs/controlflow/defs.cc

@@ -26,7 +26,7 @@ ONNX_OPERATOR_SET_SCHEMA(
     19,
     OpSchema()
         .SetDoc("If conditional")
-        .Input(0, "cond", "Condition for the if", "B")
+        .Input(0, "cond", "Condition for the if. The tensor must contain a single element.", "B")
         .Output(
             0,
             "outputs",

onnx/defs/controlflow/old.cc

@@ -25,7 +25,7 @@ ONNX_OPERATOR_SET_SCHEMA(
     16,
     OpSchema()
         .SetDoc("If conditional")
-        .Input(0, "cond", "Condition for the if", "B")
+        .Input(0, "cond", "Condition for the if. The tensor must contain a single element.", "B")
         .Output(
             0,
             "outputs",
@@ -1747,7 +1747,7 @@ ONNX_OPERATOR_SET_SCHEMA(
     1,
     OpSchema()
         .SetDoc("If conditional")
-        .Input(0, "cond", "Condition for the if", "B")
+        .Input(0, "cond", "Condition for the if. The tensor must contain a single element.", "B")
         .Output(
             0,
             "outputs",
@@ -1841,7 +1841,7 @@ ONNX_OPERATOR_SET_SCHEMA(
     11,
     OpSchema()
         .SetDoc("If conditional")
-        .Input(0, "cond", "Condition for the if", "B")
+        .Input(0, "cond", "Condition for the if. The tensor must contain a single element.", "B")
         .Output(
             0,
             "outputs",
@@ -1933,7 +1933,7 @@ ONNX_OPERATOR_SET_SCHEMA(
     13,
     OpSchema()
         .SetDoc("If conditional")
-        .Input(0, "cond", "Condition for the if", "B")
+        .Input(0, "cond", "Condition for the if. The tensor must contain a single element.", "B")
         .Output(
             0,
             "outputs",

onnx/defs/generator/defs.cc

@@ -6,6 +6,7 @@
 #include <cmath>
 
 #include "onnx/defs/function.h"
+#include "onnx/defs/generator/utils.h"
 #include "onnx/defs/schema.h"
 
 namespace ONNX_NAMESPACE {
@@ -57,115 +58,7 @@ ONNX_OPERATOR_SET_SCHEMA(
             false)
         .Output(0, "output", "Output tensor containing the same value of the provided tensor.", "T")
         .TypeConstraint("T", OpSchema::all_tensor_types_ir9(), "Constrain input and output types to all tensor types.")
-        .TypeAndShapeInferenceFunction([](InferenceContext& ctx) {
-          auto* value = ctx.getAttribute("value");
-          auto* sparse_value = ctx.getAttribute("sparse_value");
-          auto* value_int = ctx.getAttribute("value_int");
-          auto* value_ints = ctx.getAttribute("value_ints");
-          auto* value_float = ctx.getAttribute("value_float");
-          auto* value_floats = ctx.getAttribute("value_floats");
-          auto* value_string = ctx.getAttribute("value_string");
-          auto* value_strings = ctx.getAttribute("value_strings");
-
-          std::vector<bool> non_null_attr = {
-              (nullptr != value),
-              (nullptr != sparse_value),
-              (nullptr != value_int),
-              (nullptr != value_ints),
-              (nullptr != value_float),
-              (nullptr != value_floats),
-              (nullptr != value_string),
-              (nullptr != value_strings)};
-          if (std::count(non_null_attr.begin(), non_null_attr.end(), true) != 1) {
-            fail_shape_inference(
-                "One and only one of the attributes 'value', 'value_*' or 'sparse_value' must be specified for a Constant node.");
-          }
-
-          if (nullptr != value) {
-            // OpSchema::Verify check ensures that the attribute value has_t():
-            const TensorProto& tensor_proto = value->t();
-            updateOutputElemType(ctx, 0, tensor_proto.data_type());
-            updateOutputShape(ctx, 0, tensor_proto);
-            return;
-          }
-
-          if (nullptr != value_int) {
-            // OpSchema::Verify check ensures that the attribute value has_i():
-            if (!value_int->has_i()) {
-              fail_shape_inference("Attribute 'value_int' expect an integer.")
-            }
-            updateOutputElemType(ctx, 0, TensorProto::INT64);
-            updateOutputShape(ctx, 0, TensorShapeProto());
-            return;
-          }
-
-          if (nullptr != value_ints) {
-            // OpSchema::Verify check ensures that the attribute value has ints.
-            if (value_ints->ints_size() < 1) {
-              fail_shape_inference("Attribute 'value_ints' expect a list of integers.");
-            }
-            updateOutputElemType(ctx, 0, TensorProto::INT64);
-            appendDim(getOutputShape(ctx, 0), value_ints->ints_size());
-            return;
-          }
-
-          if (nullptr != value_float) {
-            // OpSchema::Verify check ensures that the attribute value has_i():
-            if (!value_float->has_f()) {
-              fail_shape_inference("Attribute 'value_float' expect a float.");
-            }
-            updateOutputElemType(ctx, 0, TensorProto::FLOAT);
-            updateOutputShape(ctx, 0, TensorShapeProto());
-            return;
-          }
-
-          if (nullptr != value_floats) {
-            // OpSchema::Verify check ensures that the attribute value has ints.
-            if (value_floats->floats_size() < 1) {
-              fail_shape_inference("Attribute 'value_floats' expect a list of floats.");
-            }
-            updateOutputElemType(ctx, 0, TensorProto::FLOAT);
-            appendDim(getOutputShape(ctx, 0), value_floats->floats_size());
-            return;
-          }
-
-          if (nullptr != value_string) {
-            // OpSchema::Verify check ensures that the attribute value has_i():
-            if (!value_string->has_s()) {
-              fail_shape_inference("Attribute 'value_string' expect a string.");
-            }
-            updateOutputElemType(ctx, 0, TensorProto::STRING);
-            updateOutputShape(ctx, 0, TensorShapeProto());
-            return;
-          }
-
-          if (nullptr != value_strings) {
-            // OpSchema::Verify check ensures that the attribute value has ints.
-            if (value_strings->strings_size() < 1) {
-              fail_shape_inference("Attribute 'value_strings' expect a list of strings.");
-            }
-            updateOutputElemType(ctx, 0, TensorProto::STRING);
-            appendDim(getOutputShape(ctx, 0), value_strings->strings_size());
-            return;
-          }
-
-          if (nullptr != sparse_value) {
-            // OpSchema::Verify check ensures that the attribute value
-            // has_sparse_tensor():
-            const SparseTensorProto& sparse = sparse_value->sparse_tensor();
-            // checker.cc::check_sparse_tensor checks that the sparse-value is
-            // well-formed
-            updateOutputElemType(ctx, 0, sparse.values().data_type());
-            auto* output_shape = getOutputShape(ctx, 0);
-            for (int i = 0; i < sparse.dims_size(); ++i)
-              appendDim(output_shape, sparse.dims(i));
-            return;
-          }
-
-          fail_shape_inference(
-              "TypeAndShapeInferenceFunction implementation incomplete: "
-              "this line should never be reached.");
-        }));
+        .TypeAndShapeInferenceFunction(ConstantOpInference));
 
 static const char* ConstantOfShape_ver20_doc = R"DOC(
 Generate a tensor with given value and shape.